Photometric and spectroscopic evolution of supernova SN ......Type Ia supernova, SN 2009an, over the...

MNRAS 430, 869–887 (2013) doi:10.1093/mnras/sts609

Photometric and spectroscopic evolution of supernova SN 2009an:another case of a transitional Type Ia event

D. K. Sahu,‹ G. C. Anupama‹ and P. AntoIndian Institute of Astrophysics, Koramangala, Bangalore 560 034, India

Accepted 2012 December 11. Received 2012 December 6; in original form 2012 May 29

ABSTRACTWe present optical UBVRI photometry and medium-resolution spectroscopy of a transitionalType Ia supernova, SN 2009an, over the period −6 to ∼+150 d from the B maximum. With a�m15(B) = 1.514 ± 0.132, SN 2009an declines faster than normal Type Ia events, but slowerthan the fast-declining, low-luminosity 1991bg-like events. The B-band absolute magnitude atmaximum is −19.02 ± 0.20. The peak bolometric luminosity indicates that 0.41 M� of 56Niwas synthesized during the explosion. The pre-maximum and early post-maximum spectralevolution of SN 2009an is very similar to that in the transitional Type Ia SN 2004eo. High-velocity features in the Ca II near-infrared triplet are seen during the early phases. Similar tothe other few objects belonging to this class, SN 2009an exhibits a higher value (∼0.4) of theSi II line ratio R(Si II). The velocity gradient of the Si II 6355 Å line in the post-maximum epoch(v̇ = 60 km s−1 d−1) is at the boundary between the low-velocity-gradient and high-velocity-gradient groups.

Key words: techniques: photometric – techniques: spectroscopic – supernovae: general –supernovae: individual: SN 2009an.

1 IN T RO D U C T I O N

Type Ia supernovae (SNe Ia) are the result of the thermonuclearexplosion of CO white dwarfs with mass close to the Chan-drasekhar limit (M ∼ Mch). Because of their almost identical pro-genitor systems, SNe Ia were thought to belong to a homoge-neous class. In a volume-limited sample, the majority of SNe Ia(∼70 per cent) belong to a fairly homogeneous class termed ‘nor-mal’ Type Ia; ∼9 per cent belong to the class of overluminous SN1991T-like objects; ∼15 per cent belong to the class of underlumi-nous SN 1991bg-like objects; and ∼5 per cent belong to the classof peculiar SN 2002cx-like objects (Li et al. 2011). The high qual-ity of observational data accumulated during the last decade hasunambiguously demonstrated that, photometrically, SNe Ia have awide range of light-curve-shape parameters, for example �m15(B)(Phillips 1993), stretch factor ‘s’ (Goldhaber et al. 2001), luminosityand mass of 56Ni synthesized in the explosion (Contardo et al. 2000).Similarly, spectroscopically there is a wide range in the expansionvelocity and its gradients (Benetti et al. 2005, Blondin et al. 2012),and in the ratio of the depth of some prominent lines, for exampleR(Si II) (Nugent et al. 1995; Blondin et al. 2012). They are, however,considered as standard candles, thanks to the empirical relationshipsbetween peak luminosities and the shape of the light curve (Phillips1993; Hamuy et al. 1996; Riess, Pree & Kirshner 1996; Phillips et al.

� E-mail: [email protected] (DKS); [email protected] (GCA)

1999; Goldhaber et al. 2001; Wang et al. 2003). There are more re-cent methods, for example SALT (Guy et al. 2005) and SALT2(Silverman et al. 2012), which measure a parametrization of thelight-curve width (x1), the supernova colour (c) and the apparentB-band magnitude at maximum light (mB). There are some spec-troscopic luminosity indicators also, such as the ratio of the depthof two absorptions at 5800 and 6100 Å (Nugent et al. 1995; Benettiet al. 2005), which is found to anti-correlate with the absolute mag-nitude, and the equivalent widths of lines due to intermediate-masselements such as Si II λ6355, which are found to be larger-valued forintermediate decliners and smaller for bright and faint objects. TheO I λ7773 line is found to be strong in fainter supernovae and tendsto become weaker with increasing luminosity (Hachinger, Mazzali& Benetti 2006). Recently, Bronder et al. (2008) found that theequivalent width of Si II λ4130 is anti-correlated with the peak B-band absolute magnitude. Extending to the light-curve stretch factor‘s’, Walker et al. (2011) and Silverman et al. (2012) showed that thehigh-stretch (brighter) objects have a lower Si II λ4130 equivalentwidth. Foley, Sanders & Kirshner (2011) found that the maximumlight velocity of Si II λ6355 and Ca II H&K, and the maximumlight pseudo-equivalent-width of Si II λ6355 are strongly correlatedwith the intrinsic (B − V) colour at peak brightness. However,Blondin et al. (2012) reported a weaker and less significant corre-lation between the Si II velocity and intrinsic (B − V) colour, whileMaguire et al. (2012) and Silverman et al. (2012) found no corre-lation between the Ca II H&K velocity and the intrinsic (B − V)colour.

C© 2013 The AuthorsPublished by Oxford University Press on behalf of the Royal Astronomical Society

by guest on Novem

ber 10, 2016http://m

nras.oxfordjournals.org/D

ownloaded from

http://mnras.oxfordjournals.org/

870 D. K. Sahu, G. C. Anupama and P. Anto

In addition to the basic explosion characteristics that determinethe properties of SNe Ia, the environment and the type of host galaxyalso influence the observed properties. The luminosity of a SN Iadepends on the morphological type of the host galaxy. SNe Ia hostedby early-type E/SO galaxies are fainter than those in spiral galax-ies, indicating that brighter events occur preferentially in youngerstellar environments (Reindl et al. 2005; Hamuy et al. 2000). Thisdependence led Mannucci et al. (2005) to propose the existenceof two populations of progenitors: (i) those found in late-typestar-forming spirals and associated with a relatively younger stellarpopulation, and (ii) those related to old stars in early-type E/SOgalaxies. This is also expressed in terms of the difference in thedelay times for progenitor formation to explosion in the two envi-ronments. From the observed bimodal distribution of delay time forSNe Ia, it is expected that the bright SNe Ia events are preferentiallyassociated with the ‘prompt’ population, which is more commonin spiral and irregular galaxies, while the faint and rapidly decay-ing objects are more easily found in the ‘tardy’ component, whichis more common in early-type galaxies (Mannucci, Della Valle &Panagia 2006).

SNe Ia with a relatively high �m15(B), relatively low ejectavelocity, moderate velocity gradient and peculiar evolution of theR(Si II) line ratio have been termed non-standard or transitionalType Ia events (Pastorello et al. 2007). They have a fainter ab-solute B maximum and produce less 56Ni than normal SNe Ia.They are supposed to bridge the gap between the ‘normal’ andthe fainter SN 1991bg-like events. This is a rare class, with onlya few well-studied objects, namely SN 2003hv (Leloudas et al.2009), SN 2004eo (Pastorello et al. 2007), SN 1992A (Kirshneret al. 1993) and SN 1989B (Wells et al. 1994). Detailed studies ofmore objects belonging to this class are required in order to gainan understanding of the properties of these transitional types ofSNe Ia.

Supernova SN 2009an was discovered by Cortitini & Antonellini(2009) on 2009 February 27.93 at an unfiltered magnitude of ∼15.4in the barred spiral galaxy NGC 4332. An independent discovery ofSN 2009an at magnitude ∼15.6 was reported by Kehusmaa (2009).Based on a spectrum obtained on 2009 February 28, Challis (2009)classified SN 2009an as a normal Type Ia supernova, about 5 dbefore maximum. Yamanaka & Arai (2009) and Vinko et al. (2009)reported the presence of a high-velocity component in the Ca II

infrared triplet at ∼22 500 km s−1, in the early spectra obtainedaround March 1.5.

Detailed photometric and spectroscopic observations, coveringthe early as well as the late phase of SN 2009an, are presented inthis paper.

2 O B S E RVAT I O N S A N D DATA R E D U C T I O N

2.1 Photometry

The monitoring of SN 2009an with the 2-m Himalayan ChandraTelescope (HCT) of the Indian Astronomical Observatory (IAO),Hanle, India, began on 2009 March 1 (JD 245 4892.35), soon afterits discovery, and continued until 2009 August 3 (JD 245 5047.13).It was monitored in the Bessell UBVRI bands, using the HimalayanFaint Object Spectrograph Camera (HFOSC) installed on the HCT.The central 2k × 2k pixels of the 2k × 4k pixels SITe CCD chipwere used for imaging. With a plate scale of 0.296 arcsec pixel−1,it covers a region of 10 × 10 arcmin2 of the sky. Standard starfields pg0918+029, pg1047+003, pg1323-086, pg0942-029 and

pg1633+099 from the list of Landolt (1992) were observed on thenights of 2009 March 15, 30 and 31 under photometric conditionsand were used to calibrate a sequence of secondary standards in thesupernova field.

The images were bias-subtracted and flat-field-corrected usingvarious tasks available within IRAF. Aperture photometry was per-formed on the standard fields, with an aperture 3–4 times the fullwidth at half-maximum (FWHM) of the stellar profile, determinedusing the aperture growth curve. The average value of atmosphericextinction for the site and the average colour terms for the systemwere used to determine the photometric zero-points on calibrationnights. A sequence of secondary standards was calibrated using theaverage colour terms and the estimated zero-points. The secondarystandards are marked in Fig. 1. The UBVRI magnitudes of the se-quence of secondary standards in the supernova field, averaged overthree nights, are listed in Table 1. The errors in the magnitudes arethe standard deviation of the standard magnitudes obtained on thethree photometric nights.

The magnitudes of the supernova and secondary standards wereestimated using the profile-fitting technique, with a fitting radiusclose to the FWHM of the stellar profile. The difference betweenthe aperture and profile-fitting magnitudes was obtained using brightstars in the supernova field, and this difference was applied to the su-pernova magnitude. The standard magnitude of the supernova wasobtained differentially with respect to the secondary standards cali-brated in the supernova field. The estimated supernova magnitudesin the U, B, V, R and I bands are listed in Table 2. The errors quotedin the magnitude table were estimated by adding in quadrature theerrors associated with the nightly photometric zero-point, the errorassociated with the aperture corrections, and the fitting errors ascomputed by IRAF.

2.2 Spectroscopy

Spectroscopic observations of SN 2009an were made on 20 epochsduring the period 2009 March 1 (JD 245 4892.29) to June 23 (JD245 5006.18). The spectra were obtained using grisms Gr#7 (wave-length range 3500–7800 Å) and Gr#8 (wavelength range 5200–9250 Å) available with HFOSC. The log of spectroscopic obser-vations is given in Table 3. Arc-lamp spectra of FeNe and FeArwere obtained for wavelength calibration. Spectrophotometric stan-dards Feige 34, Hz 44 and Feige 110, observed with the samesetup as the supernova, were used to correct for the instrumentalresponse.

Spectroscopic data reduction was carried out in the standard man-ner using the tasks available in IRAF. The two-dimensional imageswere bias-corrected and flat-fielded. The one-dimensional spectrumwas extracted using the optimal extraction method. The extractedspectra were wavelength-calibrated using the FeNe (for Grism #8)or the FeAr (for Grism #7) arc spectra obtained soon after the su-pernova spectrum. The wavelength calibration was checked usingthe night sky lines. The spectra were corrected for the continuumatmospheric extinction using the mean extinction curve for the siteand then corrected for instrumental response using the spectra ofthe spectrophotometric standards observed on the same night. Onthe nights when the spectrophotometric standards could not be ob-served, spectra obtained during nearby nights were used. The spec-tra in the two different regions were combined and scaled to aweighted mean to give the final spectrum on a relative flux scale,which was then brought to an absolute flux scale using the UBVRImagnitudes.

by guest on Novem



ownloaded from


SN 2009an: a transitional Type Ia event 871

Figure 1. Identification chart for SN 2009an. The stars used as local standards are marked as numbers 1–9. South is up and east is to the right. The field ofview is 10 × 10 arcmin2.

Table 1. Magnitudes for the sequence of secondary standard stars in the field of SN 2009an. The star IDs are as markedin Fig. 1.

ID U B V R I

1 15.168 ± 0.036 14.894 ± 0.010 14.123 ± 0.001 13.659 ± 0.007 13.243 ± 0.0202 16.309 ± 0.040 15.411 ± 0.005 14.383 ± 0.002 13.825 ± 0.002 13.347 ± 0.0133 15.145 ± 0.028 14.991 ± 0.013 14.330 ± 0.007 13.964 ± 0.002 13.617 ± 0.0104 16.760 ± 0.013 16.770 ± 0.013 16.154 ± 0.007 15.805 ± 0.014 15.452 ± 0.0215 18.223 ± 0.005 17.287 ± 0.003 16.280 ± 0.004 15.693 ± 0.007 15.194 ± 0.0166 17.371 ± 0.041 16.791 ± 0.010 15.961 ± 0.003 15.483 ± 0.007 15.088 ± 0.0137 18.835 ± 0.035 17.565 ± 0.010 16.335 ± 0.006 15.604 ± 0.015 14.927 ± 0.0298 18.129 ± 0.030 17.979 ± 0.004 17.272 ± 0.007 16.868 ± 0.008 16.471 ± 0.0019 16.520 ± 0.006 16.468 ± 0.004 15.821 ± 0.002 15.479 ± 0.010 15.109 ± 0.015

3 LI G H T C U RV E A N D C O L O U R C U RV E S

The UBVRI light curves of SN 2009an are presented in Fig. 2. Thegood photometric coverage of SN 2009an allows us to estimate thelight-curve parameters, namely epoch of maximum light in vari-ous bands, peak magnitudes, decline-rate parameter soon after thepeak �m15, and the late-phase decline rate. The date of maximum

and peak magnitudes were estimated by fitting a low-order poly-nomial to the points around maximum. Our estimates of the datesof maximum, apparent maximum magnitude and the decline-rateparameters are listed in Table 4. The light-curve fit indicates thatthe supernova reached a B-maximum light of 14.547 ± 0.025 magon JD 245 4898.5 ± 1. The decline-rate parameter in the B-band,�m15(B), is estimated as 1.514 ± 0.132. The maximum in the V and

by guest on Novem



ownloaded from



Table 2. Photometric observations of SN 2009an.

Date JD Phase∗ U B V R I245 0000+ (d)

01/03/2009 4892.35 −6.15 14.837 ± 0.029 15.067 ± 0.022 15.062 ± 0.008 14.893 ± 0.005 15.030 ± 0.01402/03/2009 4893.33 −5.17 14.659 ± 0.013 14.898 ± 0.014 14.899 ± 0.008 14.760 ± 0.007 14.863 ± 0.01303/03/2009 4894.37 −4.13 14.489 ± 0.033 14.789 ± 0.009 14.807 ± 0.015 14.657 ± 0.012 14.728 ± 0.01504/03/2009 4895.51 −2.99 14.694 ± 0.021 14.701 ± 0.012 14.551 ± 0.009 14.727 ± 0.01105/03/2009 4896.33 −2.17 14.447 ± 0.035 14.657 ± 0.025 14.659 ± 0.015 14.540 ± 0.006 14.729 ± 0.01207/03/2009 4898.11 −0.39 14.613 ± 0.092 14.585 ± 0.042 14.586 ± 0.021 14.466 ± 0.016 14.755 ± 0.01110/03/2009 4901.36 2.86 14.494 ± 0.035 14.583 ± 0.027 14.525 ± 0.022 14.406 ± 0.019 14.729 ± 0.02615/03/2009 4906.44 7.94 14.937 ± 0.029 15.082 ± 0.026 14.745 ± 0.012 14.763 ± 0.018 15.171 ± 0.01919/03/2009 4910.48 11.98 15.513 ± 0.030 15.649 ± 0.016 15.036 ± 0.015 15.099 ± 0.026 15.300 ± 0.01126/03/2009 4917.28 18.78 16.518 ± 0.009 15.422 ± 0.012 15.209 ± 0.019 15.143 ± 0.01527/03/2009 4918.25 19.75 16.642 ± 0.008 16.616 ± 0.009 15.473 ± 0.015 15.232 ± 0.012 15.148 ± 0.03829/03/2009 4920.42 21.92 16.827 ± 0.010 15.577 ± 0.006 15.261 ± 0.011 15.100 ± 0.00630/03/2009 4921.18 22.68 16.971 ± 0.010 16.891 ± 0.006 15.633 ± 0.006 15.297 ± 0.007 15.120 ± 0.01031/03/2009 4922.19 23.69 17.030 ± 0.017 17.010 ± 0.023 15.714 ± 0.008 15.358 ± 0.010 15.123 ± 0.01403/04/2009 4925.10 26.60 17.294 ± 0.021 15.945 ± 0.040 15.544 ± 0.018 15.323 ± 0.02407/04/2009 4929.12 30.62 17.620 ± 0.043 17.490 ± 0.017 16.280 ± 0.012 15.911 ± 0.017 15.619 ± 0.03110/04/2009 4932.19 33.69 17.596 ± 0.013 16.422 ± 0.015 16.102 ± 0.008 15.862 ± 0.02317/04/2009 4939.15 40.65 17.821 ± 0.025 17.820 ± 0.023 16.669 ± 0.007 16.401 ± 0.008 16.270 ± 0.01024/04/2009 4946.13 47.63 17.963 ± 0.024 17.860 ± 0.016 16.879 ± 0.019 16.677 ± 0.008 16.594 ± 0.01828/04/2009 4950.21 51.71 17.975 ± 0.015 17.945 ± 0.013 16.958 ± 0.009 16.763 ± 0.008 16.747 ± 0.01029/04/2009 4951.39 52.89 17.968 ± 0.023 16.996 ± 0.021 16.819 ± 0.018 16.838 ± 0.02107/05/2009 4959.22 60.72 18.145 ± 0.034 17.996 ± 0.031 17.215 ± 0.016 17.087 ± 0.013 17.136 ± 0.02612/05/2009 4964.35 65.85 18.308 ± 0.030 18.125 ± 0.025 17.356 ± 0.013 17.298 ± 0.024 17.371 ± 0.02521/05/2009 4973.19 74.69 18.516 ± 0.018 18.295 ± 0.021 17.597 ± 0.009 17.549 ± 0.01022/05/2009 4974.17 75.67 18.568 ± 0.027 18.255 ± 0.015 17.646 ± 0.011 17.599 ± 0.013 17.699 ± 0.02104/06/2009 4987.20 88.70 18.486 ± 0.056 17.966 ± 0.047 17.829 ± 0.053 18.139 ± 0.05206/06/2009 4989.24 90.74 18.095 ± 0.027 18.024 ± 0.02510/06/2009 4993.13 94.63 18.537 ± 0.054 18.115 ± 0.038 18.180 ± 0.019 18.290 ± 0.02723/06/2009 5006.14 107.64 18.744 ± 0.018 18.413 ± 0.021 18.524 ± 0.027 18.529 ± 0.04509/07/2009 5022.14 123.64 18.727 ± 0.012 18.965 ± 0.019 19.095 ± 0.06122/07/2009 5035.18 136.68 19.250 ± 0.018 18.973 ± 0.016 19.149 ± 0.03003/08/2009 5047.13 148.63 19.262 ± 0.027 19.430 ± 0.025

∗Observed phase with respect to the epoch of B maximum: JD 245 4898.5.

Table 3. Log of spectroscopic observations of SN 2009an.

Date JD Phase∗ Range245 0000+ (d) Å

01/03/2009 4892.29 −6.21 3500-7800; 5200-925002/03/2009 4893.35 −5.15 3500-7800; 5200-925003/03/2009 4894.41 −4.09 3500-7800; 5200-925006/03/2009 4897.31 −1.19 3500-7800; 5200-925010/03/2009 4901.30 +2.80 3500-7800; 5200-925015/03/2009 4906.49 +7.99 3500-7800; 5200-925027/03/2009 4918.20 +19.70 3500-7800; 5200-925030/03/2009 4921.31 +22.81 3500-7800; 5200-925003/04/2009 4925.42 +26.92 3500-7800; 5200-925007/04/2009 4929.21 +30.71 3500-7800; 5200-925010/04/2009 4932.40 +32.90 3500-7800; 5200-925015/04/2009 4937.29 +38.79 3500-7800; 5200-925024/04/2009 4946.26 +47.76 3500-7800; 5200-925028/04/2009 4950.27 +51.77 3500-7800; 5200-925029/04/2009 4951.32 +52.82 3500-7800; 5200-925007/05/2009 4959.25 +60.75 3500-7800; 5200-925012/05/2009 4964.30 +65.80 3500-7800; 5200-925021/05/2009 4973.23 +74.73 3500-7800; 5200-925010/06/2009 4993.18 +94.68 3500-7800; 5200-925023/06/2009 5006.18 +107.68 3500-7800; 5200-9250

∗Observed phase with respect to the epoch of B maximum:JD 245 4898.5.

R-bands occurs ∼2 d after the maximum in the B-band, whereas inthe U and I-bands the maximum occurs ∼0.5 and ∼1.5 d before theB maximum, respectively. The tendency for the I-band maximumto occur before the B-band maximum has been observed in otherSNe Ia (Anupama, Sahu & Jessy 2005; Benetti et al. 2004). The Rand I light curves each show a secondary maximum, occurring ∼20and ∼25 d after the first maximum, respectively. The secondarymaxima in the R and I-bands are ∼0.82 and ∼0.38 mag fainter thanthe first maximum.

The absolute light curve of SN 2009an in the R-band has beencompared with those of other well-studied SNe Ia, with different val-ues of the decline-rate parameter �m15(B): SN 2005hk [�m15(B) =1.68 ± 0.05, E(B − V) = 0.11, μ = 33.46; Sahu et al. (2008)], SN2004eo [�m15(B) = 1.45 ± 0.04, E(B − V) = 0.109, μ = 34.12;Pastorello et al. (2007)], SN 2003hv [�m15(B) = 1.61 ± 0.02,E(B − V) = 0.016, μ = 31.37; Leloudas et al. (2009)], SN 2003du[�m15(B) = 1.04 ± 0.04, E(B − V) = 0.01, μ = 32.89; Anupamaet al. (2005)], SN 1991T [�m15(B) = 0.95 ± 0.05, E(B − V) = 0.13,μ = 30.72; Lira et al. (1998), Phillips et al. (1992)] and SN 1991bg[�m15(B) = 1.93 ± 0.08, E(B − V) = 0.05, μ = 31.09; Filippenkoet al. (1992), Turatto et al. (1996)]. All light curves have been nor-malized to the epoch of B-band maximum for each supernova andare plotted in Fig. 3.

From Fig. 3 it is clear that in the R-band SN 2009an is more than1 mag brighter than the peculiar SN 2005hk and the underluminousSN 1991bg, fainter than SNe 1991T, 2003du and 2004eo, and has a

by guest on Novem



ownloaded from



Figure 2. UBVRI light curves of SN 2009an. The light curves have been shifted by the amount indicated in the legend.

Table 4. Photometric parameters of SN 2009an.

Data U B V R I

Epoch of max∗ 4897.9 ± 1.0 4898.5 ± 1.0 4900.9 ± 0.6 4900.4 ± 0.5 4896.1 ± 0.5Magnitude at max 14.383 ± 0.030 14.547 ± 0.025 14.530 ± 0.013 14.429 ± 0.017 14.717 ± 0.010�m15 1.519 1.514 0.826 0.757 0.581Decline rate at t > 60 d†, + 2.70 1.589 2.320 2.775 3.00

Colours at B max U − B B − V V − R R − I−0.105 ± 0.043 0.065 ± 0.032 0.071 ± 0.034 −0.288 ± 0.035

∗JD 245 0000+† mag (100 d)−1

+Days since B maximum

similar brightness to SN 2003hv. A prominent secondary maximumis seen that is absent in the light curves of the peculiar and underlu-minous Type Ia objects. Although in the early phase the R-band lightcurves of the transitional objects SNe 2009an, 2003hv and 2004eoare similar to those of SNe 1991T and 2003du, in the late phase(>60 d after B maximum) they differ considerably. The late-phasedecline rate of the transitional objects is found to be faster (declinerate ∼2.45–3.04 mag/100 d) than those of SN 2003du (decline rate1.60 mag/100 d) and SN 1991T (decline rate 1.98 mag/100 d). Asimilar trend is also seen in the late-phase decline rates in the B andV-band light curves. The B-band decline rate for transitional ob-jects is 1.58–1.76 mag/100d, as compared with the ∼1.4 mag/100 ddecline rate in SNe 1991T and 2003du. In the V-band, the de-cline rate is 2.06–2.58 mag/100 d for the transitional objects, whileit is 1.70 mag/100 d for SN 1991T and 1.53 mag/100 d for SN2003du.

The de-reddened (U − B), (B − V), (V − R) and (R − I) colourcurves of SN 2009an are plotted in Fig. 4. A reddening value ofE(B − V) = 0.108 is used (see Section 4.1). For the purpose ofcomparison, colour curves of SN 2003hv (Leloudas et al. 2009), SN2003du (Anupama et al. 2005) and SN1991T (Lira et al. 1998) arealso plotted in the same figure. The colours of the supernovae usedhere for comparison are reddening-corrected using the Cardelli,Clayton & Mathis (1989) extinction law and the E(B − V) valuesreported earlier. The (U − B), (B − V), (V − R) and (R − I) colourevolution of SN 2009an is very similar to the colour evolution of theother normal Type Ia and SN 1991T-like objects used in comparison.The colours at the B-band maximum are tabulated in Table 4.

The transitional Type Ia supernovae 2009an, 2003hv and 2004eoare all found to be intermediate decliners, with �m15(B) ∼ 1.50–1.85. They have properties that are very different from the fast-declining, underluminous SN 1991bg-like objects. The I-band

by guest on Novem



ownloaded from



Figure 3. Absolute R-band light curves of SN 2009an, plotted along with those of SNe 2005hk, 2004eo, 2003du, 2003hv, 1991T and 1991bg for comparison.The reddening and distance modulus used for estimating the absolute R magnitude are noted in the text.

Figure 4. The (U − B), (B − V), (V − R) and (R − I) colour curves ofSN 2009an. Also plotted are the colour curves of SNe 2003hv, 2003duand 1991T, for the purpose of comparison. The colour curves have beencorrected for reddening as noted in the text.

maximum is delayed with respect to the B-band maximum by afew days in the case of the fast decliners. Furthermore, these ob-jects do not show a prominent secondary maximum like that seen inthe normal SNe Ia. The SN 1991bg-like events have an unusuallyred (B − V) colour at early epochs, and they reach their reddestcolour earlier than normal SNe Ia (Modjaz et al. 2001). In con-trast to this, the intermediate decliners (transitional Type Ia) havedouble-peaked I-band light curves, and the first I-band maximumprecedes that of the B-band by a few days, as in the case of nor-mal SNe Ia. The colour evolution is also similar to that of normalSNe Ia. Thus, although somewhat underluminous compared withthe normal SNe Ia, the transitional Type Ia events are photometri-cally similar to the normal SNe Ia rather than to the underluminousones.

4 R E D D E N I N G A N D D I S TA N C E M O D U L U S

4.1 Reddening estimate

The reddening experienced by a supernova can be estimated bycomparing the observed broad-band colours of the supernova atdifferent phases with the expected colours in the absence of anyreddening as suggested by Lira et al. (1995), Phillips et al. (1999),Altavilla et al. (2004) and more recently by Folatelli et al. (2010).Lira et al. (1995) and Folatelli et al. (2010) used the uniform (B −V) colour evolution of unreddened SNe Ia between 30 and 90 d aftermaximum light, whereas Altavilla et al. (2004) and Phillips et al.(1999) used colours at maximum to estimate the reddening. Thetotal reddening estimated using the Lira et al. (1995) and Folatelli

by guest on Novem



ownloaded from



et al. (2010) methods are E(B − V) = 0.108 ± 0.044 and 0.147 ±0.047. However, the Phillips et al. (1999) and Altavilla et al. (2004)methods, which depend on the initial decline rate of the super-nova and use colour at maximum, give significantly smaller valuesfor the total reddening, namely E(B − V) = 0.035 ± 0.049 and0.011 ± 0.08.

An independent way of estimating reddening is through the equiv-alent width of the Na ID lines present in the spectra of the super-nova. Early-phase spectra of SN 2009an exhibit a narrow Na IDabsorption line in the rest-frame of the host galaxy, with an aver-age equivalent width EW = 0.56 ± 0.16 Å. Using the relationshipin Turatto, Benetti & Cappellaro (2003), we estimate E(B − V) =0.089 ± 0.025. The reddening within our Galaxy in the directionof SN 2009an is E(B − V)Gal = 0.019 (Schlegel, Finkbeiner &Davis 1998). Including the Galactic component of reddening E(B −V)Gal = 0.019, the total reddening experienced by SN 2009an isestimated to be E(B − V)total = 0.108. This value of the total red-dening is in good agreement with the estimate made using the Lirarelation. Hence, an E(B − V)total = 0.108 is used for extinctioncorrection.

4.2 The absolute magnitude of SN 2009an

NGC 4332, the host galaxy of SN 2009an, has a radial veloc-ity corrected for Local Group infall into the Virgo cluster vVir =3029 km s−1 (LEDA). Assuming H0 = 72 km s−1 Mpc−1 (Freedmanet al. 2001), we derive a distance d = 42.07 Mpc and a distance mod-ulus μ = 33.12 ± 0.15 mag. By using the total reddening estimatedas E(B − V) = 0.108 mag (see Section 4.1) and the Cardelli et al.(1989) extinction law, we obtain Mmax

U = −19.26 ± 0.20, MmaxB =

−19.02 ± 0.20, MmaxV = −18.93 ± 0.20, Mmax

R = −18.94 ± 0.20and Mmax

I = −18.56 ± 0.20.There are alternative ways to estimate the peak absolute magni-

tude of SNe Ia. The peak absolute magnitude of SNe Ia is knownto correlate well with the initial decline-rate parameter �m15(B).There are various calibrations available for the relation betweenpeak absolute magnitude and �m15(B) (Hamuy et al. 1996; Phillipset al. 1999; Altavilla et al. 2004; Reindl et al. 2005; Prieto, Rest &Suntzeff 2006). Furthermore, the intrinsic (B − V) colour at 12 d af-ter maximum light (�C12) is correlated with the absolute magnitude(Wang et al. 2005). Phillips et al. (1999) have shown that, becauseof the rapid spectral evolution of SNe Ia and the small dependenceof the reddening on the colour of the source, the observed �m15(B)must be corrected for reddening. The observed value of �m15(B)for SN 2009an translates to a reddening-corrected �m15(B)true =1.525. The peak absolute magnitudes of SN 2009an in variousbands using these calibrations are estimated and are reported inTable 5. Average values of the peak absolute magnitudes are also

reported in the same table. These average values are consistent withthe estimates made independently using the radial velocity mea-surement of the host galaxy. The estimated peak magnitudes showthat in the B-band SN 2009an is fainter than normal SNe Ia by ∼0.5mag, whereas the underluminous SN 1991bg-like objects are fainterby ∼1–2 mag than the normal Type Ia events. Furthermore, the ex-istence of secondary maxima in the R- and I-band light curvesindicates that SN 2009an is not a SN 1991bg-like underluminousobject.

The time (in days) taken by a SN Ia to reach maximum lightfrom its explosion is known as the rise time tr. The rise time is wellcorrelated with the the peak luminosity and post-maximum declinein the sense that supernovae with brighter peak luminosities andbroader light curves have longer rise times in the B-band. Preciseknowledge of the rise time provides constraints on the models ofSNe Ia (Riess et al. 1999; Fink et al. 2010). Pskovskii (1984) hasshown that the rise time can be estimated using the post-maximumdecline rate β (mag/100 d):

tr = 13 + 0.7β,

where β is closely related to �m15(B). For SN 2009an, with�m15(B)true = 1.525, we estimate a rise time of ∼20.1 ± 0.6 d,which is marginally longer than the rise time to the B maximum of19.5 ± 0.2 d for a typical SN Ia with �m15(B) = 1.1 mag and apeak MV = −19.45 (Riess et al. 1999). However, owing to the lackof photometry at very early times, we did not try to estimate the risetime using the method proposed by Riess et al. (1999).

5 TH E B O L O M E T R I C L I G H T C U RV E

The bolometric light curve of SN 2009an is estimated using thebroad-band UBVRI magnitudes presented here. In the early phase,for missing bands, magnitudes are interpolated using the points ad-jacent in time. Our U-band observations extend up to ∼75 d aftermaximum light, and we have not included the contribution of theU-band for estimating bolometric flux beyond 75 d. During the earlyphase, ∼80 per cent of the bolometric flux from a SN Ia emergesin the optical region (Suntzeff 1996), and hence the integrated fluxover the U, B, V, R and I bands gives a good estimate of the bolomet-ric flux. The magnitudes are corrected for reddening E(B − V) =0.108 (see Section 4.1) and the Cardelli et al. (1989) extinction law.The de-reddened magnitudes are then converted to monochromaticfluxes using the zero-points from Bessell, Castelli & Plez (1998).The bolometric fluxes are derived by fitting a spline curve to theU, B, V, R and I fluxes and integrating over the wavelength range3100 to 10 600 Å. The bolometric luminosity is calculated using adistance of 42.07 Mpc. The bolometric light curve of SN 2009anis plotted in Fig. 5. For the purpose of comparison, the bolometric

Table 5. Peak absolute magnitude of SN 2009an obtained using the correlationbetween �m15(B) and absolute magnitude, using various calibrations. All absolutemagnitudes have been scaled to H0 = 72 km s−1 Mpc−1.

Method MB, max MV, max MI, max

Hamuy et al. (1996) −18.703 ± 0.102 −18.744 ± 0.094 −18.527 ± 0.082Phillips et al. (1999) −18.588 ± 0.181 −18.645 ± 0.160 −18.478 ± 0.124Altavilla et al. (2004) −18.950 ± 0.142Reindl et al. (2005) −18.935 ± 0.085 −18.880 ± 0.073 −18.732 ± 0.065Wang et al. (2005) −18.901 ± 0.183 −18.922 ± 0.195 −18.668 ± 0.133Prieto et al. (2006) −18.966 ± 0.106 −18.915 ± 0.092 −18.632 ± 0.101

Average values −18.840 ± 0.157 −18.821 ± 0.121 −18.607 ± 0.103

by guest on Novem



ownloaded from



Figure 5. Bolometric light curve of SN 2009an and other well-studied SNe Ia of different subclasses.

light curves of SNe 2004eo, 2005cf, 2003du, 1991T and 1991bgare also plotted. The bolometric light curve of SN 2009an peaksat JD 245,4898.8, with a peak bolometric luminosity of log Lbol =42.89 erg s−1. Except for the marginally fainter peak, the bolometriclight curve of SN 2009an is very similar to that of SN 2004eo. Itis fainter than the bolometric light curves of normal/luminous TypeIa supernovae SNe 2005cf, 2003du and 1991T and brighter thanSN 1991bg.

Using HST and IUE spectrophotometry for SNe 1990N and1992A, Suntzeff (1996) estimated that the flux in the UV regiondrops well below 10 per cent before maximum. Contardo et al.(2000) showed that the near-infrared (NIR) passbands add at most10 per cent at early times, and no more than about 15 per cent around80 d after B maximum. Recently, Wang et al. (2009) showed thatfor SN 2005cf the UV contribution is generally a few per cent ofthe optical flux, with a peak (∼10 per cent) around the B maximum,and later the ratio remains at a level of 3–4 per cent. Using NIRdata on SNe 2005cf, 2001el and 2004S, Wang et al. (2009) showedthat the NIR to optical flux ratio shows a sharp decrease beforemaximum, and then it rises rapidly in a linear fashion and reaches apeak (∼20 per cent) at t ∼ 30 d. The NIR contribution declines lin-early during the nebular phase, reaching <10 per cent at ∼80 d afterB maximum. The correction factor required to be applied to ob-tain the uvoir luminosity from UBVRI data alone is estimated to bearound 20 per cent, with a slightly larger fraction at t ∼ 30 d owingto the appearance of a secondary maximum in the NIR bands (Wanget al. 2009). In the case of intermediate decliners �m15(B) ∼ 1.50–1.85 like SN 2004eo, the NIR JHK bands contribute ∼16 per centto the total bolometric flux around maximum and ∼37 per cent onemonth after maximum (Pastorello et al. 2007; Taubenberger et al.2008). Hence, adding about 20 per cent to the estimated flux to

compensate for the missing bands in the UV and NIR, the peakbolometric flux is estimated to be log Lbol = 42.97 erg s−1.

Another important parameter, the mass of 56Ni synthesized in theexplosion, can be estimated using the peak bolometric flux. UsingArnett’s law (Arnett 1982) and the expression for the maximumluminosity produced by the radioactive decay of 56Ni (Stritzinger& Leibundgut 2005),

Lmax = (6.45e−tr/(8.8d) + 1.45e−tr/(111.3d)

)

× (MNi/M�

) × 1043 erg s−1,

where tr is the rise time of the bolometric light curve, and MNi is themass of 56Ni. The estimated B-band rise time of SN 2009an is 20 d.Contardo et al. (2000) found the rise time of the bolometric lightcurve to be within one day of the rise time in the B-band. Using arise time tr of 20 d and a peak bolometric luminosity of log Lbol =42.89 erg s−1, the mass of 56Ni synthesized during the explosionis estimated to be 0.41 M�. If another 20 per cent of flux is addedto account for the missing UV and NIR fluxes, the mass of 56Ni isestimated to be 0.50 M�. The other objects with an intermediatedecline rate in the B-band light curve, for example SNe 1992A,2003hv and 2004eo, have a mass of 56Ni in the same range.

6 SPECTRO SCOPI C RESULTS

6.1 Spectral evolution

The spectral evolution of SN 2009an is presented in Figs 6, 9 and 11.Fig. 6 shows the spectral evolution at the pre-maximum and aroundthe maximum phase. The prominent lines in the pre-maximumspectrum were identified following Kirshner et al. (1993). The

by guest on Novem



ownloaded from



Figure 6. Spectral evolution of SN 2009an in the pre-maximm and around maximum phase (day −6 to day +7). The phases are relative to the date of Bmaximum. The spectra have been corrected for the host-galaxy redshift and reddening. For clarity, the spectra are displaced vertically. The main features aremarked. The telluric lines have not been removed and are marked with the symbol ⊕.

post-maximum spectral evolution is plotted in Fig. 9, and the earlynebular and nebular phase spectral evolution is shown in Fig. 11.The prominent features in the spectra were identified followingMazzali et al. (2008), who modelled the spectra of SN 2004eo.

The first spectrum, observed on JD 245 4892.29, correspondsto ∼6 days before B maximum. It shows a blue continuum with awell-developed Si II 6355 Å absorption, and a weaker Si II 5972 Å.Other features in the spectrum include those due to Ca II H&K,Mg II 4481 Å, possibly blended with Si III lines, and well-developedW-shaped S II lines at 5454 and 5640 Å. Blends due to Fe II, Si II

and S II are also seen between 4000 and 5000 Å. The Ca II NIRtriplet at 8498, 8542 and 8662 Å appears to have two componentscorresponding to low and high velocities. There is a hint of O I

7774 also in the first spectrum, which becomes stronger and isseen prominently until ∼+20 d. The pre-maximum and early post-maximum spectra appear similar, with some noticeable changesduring this interval. The absorption strength of the Si II, S II andCa II lines increases during this phase. The increase in the strengthof the S II line is attributed to recombination, and to an increasedabundance as deeper layers are exposed (Mazzali et al. 2008). Thesingle absorption trough between 4800 and 5000 Å identified asFe III in the −6 d spectrum splits into two troughs in the +2 dspectrum as Fe II dominates over Fe III. A change in the shape ofthe continuum is noticed in the spectrum of +7 d. A high-velocityfeature in the Ca II NIR triplet is seen in the spectra during theinterval −6 to +7 d.

In Fig. 7, spectra of SN 2009an at t = −6 d (top panel) and att = −1 d (bottom panel) are plotted. Spectra of SN 2004eo (Pas-torello et al. 2007), SN 2003du (Anupama et al. 2005), SNe 1994D

and 1996X (Salvo et al. 2001) are also plotted in the same figure forcomparison. All SNe show the presence of the high-velocity com-ponent in the Ca II IR triplet feature. It is also seen that the spectrumof SN 2009an is very similar to that of SN 2004eo, with comparablestrength and widths of lines due to singly ionized intermediate-masselements. In both these SNe, the Si II 6355 Å line is broad, and theSi II 5972 Å absorption is stronger than in SN 2003du. The Ca II

H&K lines show a single, flat-bottomed profile, while in SN 1994Dit is split into two. The line profile around 4500 Å is also found to bedifferent. While SN 2009an (and SN 2004eo) shows a single deepabsorption, the other SNe Ia, such as SNe 1994D and 1996X, showtwo absorption troughs, which are identified as due to Mg II 4481 Åand Si III 4553, 4568 Å, with almost equal strength. We identifythe absorption in SN 2009an to be due to Mg II, with Si III beingweak or absent. The strength of the Si III line is found to be corre-lated with temperature (Pignata et al. 2008). This indicates that SN2009an has a lower photospheric temperature than SNe 1994D or1996X.

We use the parametrized spectrum synthesis code ‘SYN++’ toidentify various species in the observed spectrum and to derive otherphysical parameters. ‘SYN++’ is a rewrite of the original SYNOW code(refer to https://c3.lbl.gov/es/ for further details of SYN++) with anew structured input control file format and more complete atomicdata files. The SYNOW code assumes a spherically symmetric andhomologously expanding ejecta (v ∝ r) and resonant scattering lineformation above a sharp photosphere emitting a blackbody contin-uum. The line formation is treated using the Sobolev approximation(Sobolev 1957; Jeffery & Branch 1990). The optical depth τ of thestrongest line is a free-fitting parameter, and the optical depths of

by guest on Novem



ownloaded from



Figure 7. Pre-maximum spectrum of SN 2009an compared with those of Type Ia SNe 2004eo, 1994D, 2003du and 1996X at a similar epoch. The phasesmarked are relative to the date of the B maximum. The spectra have been corrected for the host galaxy redshift and reddening. For clarity the spectra are shiftedvertically.

the other lines of the same ion are determined assuming Boltzmannequilibrium at excitation temperature Texc. A detailed descriptionof the SYNOW code can be found in (Fisher 2000) and (Branch et al.2002)).

The −5, −1, +8 and +31 d spectra of SN 2009an have beenmodelled with the SYN++ code. The early phase observed and syn-thetic spectra are plotted in Fig. 8. The −5 d spectrum of SN 2009anfits well with a synthetic spectrum with a blackbody temperatureof 14 800 K and a photospheric velocity vphot of 11 500 km s−1. Itcontains lines of O I, Mg II, Si II, Si III, S II and Fe III with a photo-spheric velocity component only, while Ca II has both photosphericand high-velocity components (Fig. 8, top panel). The absorptionsnear 4000 and 4300 Å and the sawtooth-like line profile at 5000 Åcan be reproduced only by introducing a Fe III line at lower veloc-ity. Furthermore, the absorption at 5700 Å cannot be reproducedby only Si II, and introducing Fe III improves this fit also. The pres-ence of Fe III and Si III lines in the −5 d spectrum also indicatesa high temperature in the ejecta. The overall shapes of the −1 dspectrum and the −5 d spectrum are very similar, and the observedspectrum of −1 d (Fig. 8, middle panel) can be reproduced with thesame species as used in the −5 d spectrum, with a marginallylower blackbody temperature (∼14 300 K) and a lower veloc-ity for the high-velocity component of Ca II. The +8 d spectrumof SN 2009an indicates an evolution towards lower temperatureand lower photospheric velocity. The observed spectrum matcheswell with a synthetic spectrum with a blackbody temperature of11 500 K and a photospheric velocity vphot of 6500 km s−1 (Fig. 8,bottom panel). At 5900 Å, the absorption feature is well repro-duced by Si II, Fe III and Na I, indicating the emergence of Na I. In

the underluminous events, the introduction of Ti II improves thefit near 4200 Å. However, in the case of SN 2009an, the intro-duction of Ti II does not improve the fit, implying the absence ofTi II.

The spectral evolution during +20 to +53 d is shown in Fig. 9(top panel). During this interval, the Si II line at 5972 Å is replacedby the Na I line, which starts becoming stronger and contaminatingthe features near 5900 Å. Mazzali et al. (2008) identified the broadfeatures between 4500 and 5000 Å with Fe II, probably with somecontribution from Ti II, as is seen in the underluminous SN 1991bg-like events. The synthetic spectrum generated for day +31 is dis-played along with the observed spectrum in Fig. 9 (bottom panel).The synthetic spectrum has a blackbody temperature of 7000 K anda photospheric velocity vphot of 6000 km s−1. During this phase, theabsorption core at ∼6100 Å is reproduced by the Si II feature, andthe strengthening of the Fe II features in the adjacent region makethe absorption broader. The strong absorption at ∼5900 Å is repro-duced mainly by Na I. The post-maximum spectra of SN 2009anare similar to those of other SNe Ia, as seen from Fig. 10.

The early nebular and nebular spectra obtained during days +61to +107, shown in Fig. 11, indicate that the spectrum is dominatedby forbidden lines of singly and doubly ionized Co and Fe byday +61. Not much evolution is seen in the spectrum during thisphase. A comparison of the nebular spectra of SN 2009an withthose of other SNe Ia (Fig. 12) indicates an overall similarity in allsupernovae.

The overall spectral evolution of SN 2009an is very similar tothat of SN 2004eo, which was shown by Pastorello et al. (2007) tobe similar to that of SN 1992A, although with a lower velocity.

by guest on Novem



ownloaded from



Figure 8. Observed spectra of SN 2009an (red) at −5 d (top panel), −1 d (middle panel) and +8 d (bottom panel). The observed spectra are plotted with thesynthetic spectra (black) obtained with SYN++.

6.2 Expansion velocity of the ejecta

The expansion velocity of the ejecta was measured by fitting a Gaus-sian to the absorption trough of Ca II H&K, S II 5640 Å, Si II 6355 Åand the Ca II NIR triplet. The velocity evolution of various speciesis plotted in Fig. 13. We confirm the presence of a high-velocitycomponent in the Ca II NIR triplet during the pre-maximum andearly post-maximum phase, noted by Yamanaka & Arai (2009) andVinko et al. (2009) in the spectrum of 2009 March 1. Expansionvelocity estimates for both the high- and the low-velocity compo-nents are plotted in Fig. 13. Mazzali et al. (2005) have shown thatalmost all SNe Ia with early spectra show a high-velocity feature inthe Ca II NIR triplet, but with varying strength and duration. Theymay be interpreted as being caused either by the enhancement in theabundance of Ca in the outer region, or by density enhancementsas a result of the sweeping up of the interstellar material by thehighest-velocity supernova ejecta (Mazzali et al. 2005). Gerardyet al. (2004) interpret the presence of the high-velocity feature asan indication of interaction between the circumstellar material andthe supernova ejecta.

Based on the Si II velocity gradient between the time of maximumand either the time the Si II feature disappears or the last availablespectrum, whichever is earlier, Benetti et al. (2005) arranged SNe Iainto three classes: (i) the FAINT class consisting of faint objectswith a high velocity gradient and high Si II line ratio; (ii) the HVG

class consisting of normal SNe Ia with a high velocity gradient(>70 km s−1 d−1) and low Si II line ratio; and (iii) the LVG class witha low velocity gradient (<60–70 km s−1 d−1). This class includesboth normal and bright SNe Ia. In the case of SN 2009an, thevelocity gradient of the Si II λ 6355 Å line during the interval ∼+3to +39 d is estimated to be 60 km s−1 d−1, which is just at the upperlimit of the LVG class, similar to SNe 2004eo, 1992A and 1989B.However, Benetti et al. (2005) have also noted that for SNe 1989Band 1992A the classification is uncertain. A small variation of theparameters, still within the errors, could shift them from the LVGgroup to the HVG group, and the same holds for SN 2009an.

Recently, Blondin et al. (2012) suggested that instead of estimat-ing the velocity gradient from a least-squares fit of the measurementstaken between maximum and either the time the Si II feature disap-pears or the last available spectrum, the mean velocity decline rateshould be measured over some fixed time interval. For SN 2009an,if the measured velocities at −1.19 and +7.99 d are used, a velocitydecline rate of ∼93 km s−1 d−1 is estimated, shifting SN 2009anfrom the LVG to the HVG group.

The expansion velocity estimated using the Si II 6355 Å line isplotted in Fig. 14, with other objects belonging to the LVG and HVGgroups and objects with faster-declining light curves. In the pre-maximum phase, SN 2009an has a higher expansion velocity thanother objects with similar �m15(B). After the maximum, however,the expansion velocity is similar to that of SN 1992A. There are

by guest on Novem



ownloaded from



Figure 9. Top panel: Spectral evolution of SN 2009an during +20 to +53 d relative to the date of B maximum. The spectra have been corrected for thehost-galaxy redshift and reddening. For clarity the spectra are displaced vertically. The telluric lines have not been removed and are marked with the symbol⊕. Bottom panel: +31-d spectrum of SN 2009an (red) compared with a synthetic spectrum obtained using SYN++ (black).

some LVG supernovae that show a high expansion velocity in thepre-maximum phase, but a few days after maximum the differencein the expansion velocity among different class is not significant.

It has been suggested that the difference in the velocity gradientfor the HVG and LVG classes can come either from the nature ofthe explosion or from the different degrees of mixing of heavy ele-ments. One possibility suggested by Benetti et al. (2005) is that thesupernovae belonging to the HVG group are the result of delayeddetonation, while the explosion mechanism for the LVG group ob-jects is deflagration. The high velocity gradient in the HVG classcould arise from the efficient mixing of heavy elements to the outerejecta, whereas less efficient mixing could be the reason for the lowvelocity gradient of the LVG group. In a recent theoretical study,Maeda et al. (2010) showed that a white dwarf initially ignitedslightly away from the centre results in an asymmetric explosion.They suggest that the HVG and LVG supernovae do not have intrin-sic differences, but that the observed diversity could be an effect ofthe viewing angle of such an asymmetric explosion. The supernovaethat are viewed from the direction of the off-centre ignition appearas LVG events, while those viewed from the opposite directionappear as HVG events.

6.3 R(Si II) ratio

The ratio R(Si II) of the depths of two Si II lines at 5972 and 6355 Å,as defined by Nugent et al. (1995), is known to be correlated with the

absolute magnitude of SNe Ia. At the B-band maximum, the ratio isfound to be large for dimmer objects and smaller for brighter objects,and hence it could be used as a distance-independent spectroscopicluminosity indicator.

The R(Si II) ratio for SN 2009an is calculated using the methodprescribed by Nugent et al. (1995) during the pre-maximum andearly post-maximum phase. The temporal evolution of the R(Si II)ratio for SN 2009an along with that for other supernovae are plottedin Fig. 15. With an average pre-maximum value of 0.42 ± 0.03,no significant temporal evolution of the R(Si II) ratio is seen in thecase of SN 2009an. Benetti et al. (2005) noted that in the pre-maximum phase, HVG SNe Ia show significant temporal evolution,starting from a high value well before maximum and levelling outjust before maximum, whereas the R(Si II) ratio for the LVG SNe Iastays nearly constant. Around maximum, the R(Si II) ratios of HVGand LVG supernovae are comparable. In Fig. 15, it is clearly seenthat, in terms of the R(Si II) ratio, SNe 1989B, 1992A, 2004eo and2009an are different from all other supernovae; in fact, they liebetween the normal SNe Ia and the underluminous 1991bg-likeevents. Similar to the LVG group supernovae, the R(Si II) ratios forSNe 1989B, 1992A, 2004eo and 2009an do not show significanttemporal evolution.

The correlation between the R(Si II) ratio and the luminosityof SNe Ia was interpreted as resulting from the temperature dif-ference, which is directly related to the variation in the mass of56Ni produced in the explosion. At lower effective temperatures, a

by guest on Novem



ownloaded from



Figure 10. Comparison of the post-maximum spectrum of SN 2009an with those of Type Ia SNe 2004eo, 2003du, 1996X and 1994D, at similar epochs withrespect to the date of B maximum. The spectra have been corrected for the host-galaxy redshift and reddening. For clarity the spectra are shifted vertically.

Figure 11. Spectral evolution of SN 2009an during +61 to +107 d relative to the date of B maximum. The spectra have been corrected for the host-galaxyredshift and reddening. For clarity the spectra are displaced vertically. The telluric lines have not been removed and are marked with the symbol ⊕.

by guest on Novem



ownloaded from



Figure 12. Late-phase spectrum of SN 2009an during ∼+60 to ∼+107 d compared with those of Type Ia SNe 2004eo, 2003du, 1996X and 1994D at a similarepoch relative to B maximum. The spectra have been corrected for the host galaxy redshift and reddening. For clarity the spectra are shifted vertically.

Figure 13. Velocity evolution of SN 2009an derived using the Si II λ 6355 Å Ca II H&K, Ca II NIR and S II λ 5640 Å lines.

by guest on Novem



ownloaded from



Figure 14. Velocity evolution of SN 2009an derived using the Si II λ 6355 Å line. The evolutions for other supernovae are also plotted for comparison. Objectsbelonging to the HVG class are represented by filled symbols, the transitional Type Ia objects are marked as open starred symbols, and open symbols representthe LVG objects.

complex interaction of the lines with line blanketing from Fe II andCo II increases the apparent strength of the 5800 Å line, and athigher effective temperatures Fe III and Co III line blanketing helpsto wash out the feature at 5800 Å (Nugent et al. 1995). However,Hachinger et al. (2008) have shown that the increase of the R(Si II)ratio for luminous to dim object is caused by the increasing strengthof the Si II 5972 Å line for dim objects. In less luminous objects, Si II

5972 Å is stronger because of a rapidly increasing Si II/Si III ratio.Because of its large optical depth, the Si II 6355 Å line is saturatedand has a fairly constant strength. These authors also suggested thatas the Si II 6355 Å line is saturated, the Si II 5972 Å line may be thebest spectroscopic luminosity indicator for SNe Ia.

Branch, Dang & Baron (2006) and Branch et al. (2009) arrangedthe maximum light spectra of SNe Ia on the basis of measure-ments of W(5750) and W(6100), the (pseudo) equivalent widths ofabsorption features near 5750 Å (usually attributed to Si II λ5972)and 6100 Å (due to Si II λ6355) and the appearance of 6100 Å ab-sorption. They classified the spectra of SNe Ia into four groups –core-normal, broad-line, cool, and shallow-silicon. The core-normalsupernovae have very similar spectra at maximum light and post-maximum epochs, except for the varying strength of high-velocityCa II features. The broad-line supernovae have 6100 Å absorptionbroader and deeper than that of core-normal supernovae. Cool su-pernovae have a conspicuous absorption trough at ∼4000–4500 Ådue to Ti II, whereas those in the shallow-silicon group have smallervalues of W(5750) and W(6100). As per the original classification byBranch et al. (2009), SN 1992A was classified as a cool supernova,SN 1989B was classified as marginal cool, with no identifiable Ti II

lines, and SN 2004eo was classified as mild cool. In Section 6.1, itis shown that the synthetic spectra of SN 2009an calculated usingthe SYN++ code do not need the presence of Ti II.

The pseudo-equivalent widths W(5750) and W(6100), for a sus-bset of data used by Branch et al. (2009), after including SNe2003hv and 2009an, are plotted in Fig. 16. The five objects SNe1989B, 1992A, 2003hv, 2004eo and 2009an, termed transitionalobjects, form an easily identifiable group that nicely fills the gapbetween the cool and core-normal supernovae. It indicates a gradualtransition of properties from the cool supernovae to the core-normalsupernovae, and thus the possible presence of a continuous distri-bution of properties amongst SNe Ia of different subclasses. In fact,using a sample that includes objects with a wider range of �m15(B)between ∼0.7 and ∼2 mag, Blondin et al. (2012) have shown thatthere are no strict boundaries between the different subclasses andthat there is a continuum of properties between them.

In Fig. 17, similar to fig 3(a) of Benetti et al. (2005), the value ofthe R(Si II) ratio at maximum light for a subsample of Branch et al.(2009), including the transitional events, is plotted against �m15(B).In fig. 3(a) of Benetti et al. (2005) there is a gap between the groupcontaining faint objects and the group containing HVG and LVGobjects. Pastorello et al. (2007) have shown that SN 2004eo andother non-standard objects, for example SNe 1989B and 1992A, lieroughly in the middle of the diagram, between the faint group andthe HVG/LVG group. In Fig. 17 we have used the nomenclaturein Branch et al. (2009). It is clear from the figure that the transi-tional objects are located between the cool and core-normal objects,connecting the faint and normal bright Type Ia objects. Following

by guest on Novem



ownloaded from



Figure 15. Temporal evolution of the R(Si II) ratio of SN 2009an, together with those of other supernovae for comparison. The SN 1991bg-like underluminoussupernovae are marked as skeletal (centre connected to vertices) symbols, transitional-type events are marked as open starred symbols, filled symbols representthe HVG supernovae, and open symbols represent the LVG supernovae.

the criterion suggested by Benetti et al. (2005), the transitional ob-jects fall in the LVG group, so it will be appropriate to considerthe transitional objects as a link between the faint and the LVGSNe Ia.

7 SU M M A RY

We have presented UBVRI photometric observations of SN 2009anduring the interval ∼−6 to ∼+150 d with respect to the B-bandmaximum. The light-curve evolution of SN 2009an shows that thesupernova reached its B-band maximum on JD 245 4898.5 ± 1at 14.547 ± 0.025 mag. The light curves of SN 2009an declinesfaster than the normal SNe Ia, with a B-band decline rate parameter�m15(B) of 1.514 ± 0.132. The moderately higher value of �m15(B)for SN 2009an puts it along with SNe 2004eo and 1992A in thetransitional class, between the ‘normal’ and ‘low-luminosity SN1991bg-like objects’. The peak absolute magnitude of SN 2009anis estimated using empirical relations. With an absolute B-bandmagnitude at peak MB, max = −18.841 ± 0.157, SN 2009an is foundto be dimmer than the normal Type Ia objects. The UBVRI bolo-metric luminosity of SN 2009an is derived, and the peak bolometricluminosity log Lbol is estimated as 42.89 erg s−1; this implies that0.41 M� of 56Ni was synthesized in the explosion. However, if an-other 20 per cent flux is added to account for the missing UV andNIR bands, the mass of 56Ni is estimated to be 0.50 M�, which isvery close to the mass of 56Ni synthesized in the explosions of otherevents belonging to the transitional class.

The spectroscopic data during ∼−6 to ∼+107 d are also pre-sented. The pre-maximum spectral evolution of SN 2009an is very

similar to that of the transitional-type object SN 2004eo. High-velocity features are seen in the Ca II NIR triplet during the pre-maximum and early post-maximum phase. In the pre-maximumand early post-maximum spectra, SN 2009an has a broad Si II 6355Å line, and Si II 5972 Å is stronger than for normal Type Ia events.The post-maximum spectral evolution is similar to those of normalType Ia objects. Based on the Si II 6355 Å line velocity gradient of60 km s−1 d−1 during +3 to +39 d, SN 2009an is just at the upperlimit of the LVG class. However, if the revised definition of thevelocity decline rate is used, SN 2009an falls in the HVG class. TheR(Si II) ratio does not show any significant evolution and its averagevalue is 0.42 ± 0.03, which is between that for normal Type Ia andunderluminous SN 1991bg-like events.

A study of the properties of all the known transitional Type Iaevents indicates that their photometric properties (�m15(B), MB) andtheir spectroscopic luminosity indicator R(Si II) ratio lie betweenthose of normal Type Ia objects and the underluminous SN 1991bg-like objects. In the W(5750) versus W(6100) and R(Si II) versus�m15(B) plots, the transitional objects occupy the space betweenunderluminous 1991bg-like objects and the normal Type Ia events.

To explore whether these transitional SNe Ia bridge the gap be-tween the normal Type Ia events and the low-luminosity SN 1991bg-like events, it is desirable to have a detailed model for these objects.The light curve and spectra of SN 2004eo have been modelled byMazzali et al. (2008). These authors have shown that the spectralevolution, bolometric light curve and low 56Ni mass of SN 2004eocan be reproduced by explosion of a Chandrasekhar-mass CO whitedwarf with a lower kinetic energy (1.1 ± 0.1 × 1051 erg) than theW7 model kinetic energy (1.3 × 1051 erg), and a larger amount of

by guest on Novem



ownloaded from



Figure 16. Plot of the pseudo-equivalent width of Si II 5972 line, W(5750), versus the pseudo-equivalent width of Si II 6355 line, W(6100), for a subsample ofSNe Ia. The nomenclature is as in Branch et al. (2009).

stable Fe-group material. Mazzali et al. (2008) also indicated thatsuch low-luminosity explosions may be the result of a strong ini-tial deflagration phase, during which the star expands, and, whena delayed detonation occurs, it burns the material at low density,mostly to the intermediate mass element, which is expected to beunburned oxygen, as signature of unburned carbon is not seen in thespectra. The absence of carbon and the presence of absorption dueto oxygen in the pre-maximum and early post-maximum spectra ofSN 2009an are consistent with the explosion model proposed forSN 2004eo. The observed similarity between SN 2004eo and SN2009an in terms of their photometric and spectroscopic propertiesmay indicate that these events are the results of identical explosions,with lower kinetic energy than in the W7 model.

Even with a limited number of the transitional objects studied indetail, it appears that they fill the gap between the underluminousand the normal Type Ia events. The observed dissimilarity betweenthe transitional and normal Type Ia events may also be influencedby the host galaxy properties. Curiously enough, the host galaxiesof the known transitional objects are all group members. They areeither lenticulars or spirals, and the SN is located away from thenucleus (which could be just an observational bias). We do not,however, have any information about the host galaxy metallicity, inparticular at the SN location. To understand these events better, andto examine whether the transitional objects really bridge the gapbetween the normal Type Ia events and low-luminosity SN 1991bg-like objects, it is necessary to have a detailed study, including therobust modelling, of as many objects as possible belonging to thisclass. Upcoming surveys may discover more such objects, and, with

a sufficiently large sample, the properties of this class of objects canbe better understood.

Finally, we try to check whether the inclusion of transitionalobjects has any effect on cosmology. The peak absolute magnitudeof the transitional objects, estimated using the extinction-correctedobserved magnitude and the distance modulus, was used to estimate�m15(B), with the help of empirical relations between absolutemagnitude and �m15(B). It is found that the empirical relationsreproduce the observed �m15(B) well, implying that their inclusionmay not affect the results of SN cosmology.

AC K N OW L E D G M E N T S

We would like to thank the anonymous referee, whose commentshelped to improve the work considerably. We would like to expressour sincere thanks to Gaston Folatelli for discussion on SYN++fitting. We thank all the observers at the 2-m HCT who kindlyprovided some of their observing time for the supernova observa-tions, and Ms Sonam Arora for help with observations. This workhas made use of the NASA Astrophysics Data System and theNASA/IPAC Extragalactic Data base (NED), which is operated bythe Jet Propulsion Laboratory, California Institute of Technology,under contract with the National Aeronautics and Space Adminis-tration. IRAF is distributed by the National Optical Astronomy Obser-vatories, which are operated by the Association of Universities forResearch in Astronomy, Inc., under contract to the National ScienceFoundation.

by guest on Novem



ownloaded from



Figure 17. Plot of the R(Si II) ratio versus �m15 for a subsample of SNe Ia used in Figs 15 and 16. The nomenclature is as in Branch et al. (2009).

R E F E R E N C E S

Altavilla G. et al., 2004, MNRAS, 349, 1344Anupama G. C., Sahu D. K., Jessy J., 2005, A&A, 429, 667Arnett W. D., 1982, ApJ, 253, 785Benetti S. et al., 2004, MNRAS, 348, 261Benetti S. et al., 2005, ApJ, 623, 1011Bessell M. S., Castelli F., Plez B., 1998, A&A, 333, 231Blondin S. et al., 2012, AJ, 143, 126Branch D., Benetti S., Kasen D., Baron E., Jeffery D. J., Hatano K., 2002,

ApJ, 566, 1005Branch D. et al., 2006, PASP, 118, 560Branch D., Dang L. C., Baron E., 2009, PASP, 121, 238Bronder T. J. et al., 2008, A&A, 477, 717Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345, 245Challis P., 2009, CBET, 1709, 1Contardo G., Leibundgut B., Vecca W. D., 2000, A&A, 359, 876Cortitini G., Antonellini S., 2009, CBET, 1707, 1Filippenko A. V. et al., 1992, AJ, 104, 1543Fink M., Ropke F. K., Hillebrandt W., Seitenzahl I. R., Sim S. A., Kromer

M., 2010, A&A, 514, 53Fisher A., 2000, PhD thesis, Univ. OklahomaFolatelli G., Phillips M. M., Burns C., Contreras C., Hamuy M., Freedman

W. L., Persson S. E., Stritzinger M., 2010, AJ, 139, 120Foley R. J., Sanders N. E., Kirshner R. P., 2011, ApJ, 742, 89Freedman W. L., Madore B. F., Gibson B. K., Ferrarese L., Kelson D. D.,

Sakai S., Mould J. R., Kennicutt R. C., 2001, ApJ, 553, 47Gerardy C. L. et al., 2004, ApJ, 607, 391Goldhaber G. et al., 2001, ApJ, 558, 359Guy J., Astier P. Nobili S., Regnault N., Pain R., 2005, A&A, 443, 781Hachinger S., Mazzali P. A., Benetti S., 2006, MNRAS, 370, 299Hachinger S., Mazzali P. A., Tanaka M., Hillebrandt W., Benetti S., 2008,

MNRAS, 389, 1087

Hamuy M., Phillips M. M., Suntzeff N. B., Schommer R. A., Jose M., SmithR. C., Lire P., Aviles R., 1996, AJ, 112, 2438

Hamuy M., Trager S. C., Pinto P. A., Phillips M. M., Schommer R. A.,Ivanov V., Suntzeff N. B., 2000, AJ, 120, 1479

Jeffery D. J., Branch D., 1990, in Wheeler J. C., Piran T., Weinberg S., eds,Jerusalem Winter School for Theoretical Physics Vol. 6, Supernovae.World Scientific, Singapore, p. 149

Kehusmaa P., 2009, CBET, 1707, 1Kirshner R. P. et al., 1993, ApJ, 415, 589Landolt A. U., 1992, AJ, 104, 340Leloudas G. et al., 2009, A&A, 505, 265Li W. et al., 2011, MNRAS, 412, 1441Lira P., 1995, MSc thesis, Univ. ChileLira P., Suntzeff N. B., Phillips M. M., Hamuy M., Maza J., Schommer

R. A., Smith R. C., Wells L. A., 1998, AJ, 115, 234Maeda K. et al., 2010, Nat, 466, 82Maguire K. et al., 2012, MNRAS, 426, 2359Mannucci F., Della Valle M., Panagia N., Cappellaro E., Cresci

G., Maiolino R., Petrosian A., Turatto M., 2005, A&A, 433,807

Mannucci F., Della Valle M., Panagia N., 2006, MNRAS, 370, 773Mazzali P. A., Benetti S., Altavilla G., Blanc G., Cappellaro E., Elias-Rosa

N., Garavini G., Goobar A., 2005, ApJ, 623, L37Mazzali P. A., Sauer D. N., Pastorello A., Benetti S., Hillebrandt W., 2008,

MNRAS, 386, 1897Modjaz M., Li W., Filippenko A. V., King J. Y., Leonard D. C., Matheson

T., Treffers R. R., 2001, PASP, 113, 308Nugent P., Phillips M., Baron E., Branch D., Hauschildt P., 1995, ApJ, 455,

L147Pastorello A. et al., 2007, MNRAS, 376, 1301Phillips M. M., 1993, ApJ, 413, L105Phillips M. M., Suntzeff N. B., Hamuy M., Leibundgut B., Kirshner R. P.,

Foltz C. B., 1992, AJ, 103, 1632

by guest on Novem



ownloaded from



Phillips M. M., Lira P., Suntzeff N. B., Schommer R. A., Hamuy M., JoseM., 1999, AJ, 118, 1766

Pignata G. et al., 2008, MNRAS, 388, 971Prieto J. L., Rest A., Suntzeff N. B., 2006, ApJ, 647, 501Pskovskii Y. P., 1984, Sov. Astron., 28, 658Reindl B., Tammann G. A., Sandage A., Saha A., 2005, ApJ, 624, 532Riess A. G., Pree W. H., Kirshner R. P., 1996, ApJ, 473, 88Riess A. G. et al., 1999, AJ, 118, 2675Sahu D. K. et al., 2008, ApJ, 680, 580Salvo M. E., Cappellaro E., Mazzali P. A., Benetti S., Danziger I. J., Patat

F., Turatto M., 2001, MNRAS, 321, 254Schlegel D. J., Finkbeiner D. P., Davis M., 1998, ApJ, 500, 525Silverman J. M., Ganeshalingam M., Li W., Filippenko A. V., 2012,

MNRAS, 425, 1889Sobolev V. V., 1957, Sov. Astron., 1, 678Stritzinger M., Leibundgut B., 2005, A&A, 431, 423Suntzeff N. B., 1996, in McCray R., Wang Z., eds, Supernovae and Super-

novae Remnants. Cambridge Univ. Press, Cambridge, p. 41

Taubenberger S. et al., 2008, MNRAS, 385, 75Turatto M., Benetti S., Cappellaro E., Danziger I. J., Della Valle M., Gouiffes

C., Mazzali P. A., Patat F., 1996, MNRAS, 283, 1Turatto M., Benetti S., Cappellaro E., 2003, in Hillebrandt W., Leibundgut

B., eds, Proc. ESO-MPA-MPE Workshop, From Twilight to Highlight:The Physics of Supernovae. Springer, Berlin, p. 200

Vinko J., Marion H., Wheeler J. C., Gerardy C., 2009, CBET, 1710, 1Walker E. S. et al., 2011, MNRAS, 410, 1262Wang L., Goldhaber G., Aldering G., Perlmutter S., 2003, ApJ, 590, 944Wang X., Wang L., Zhou X., Lou Y.-Q., Li Z., 2005, ApJ, 620, L87Wang X. et al., 2009, ApJ, 697, 380Wells L. A. et al., 1994, AJ, 108, 2233Yamanaka M., Arai A., 2009, CBET, 1710, 1

This paper has been typeset from a TEX/LATEX file prepared by the author.

by guest on Novem



ownloaded from


Photometric and spectroscopic evolution of supernova SN ......Type Ia supernova, SN 2009an, over the...

Documents

Transcript of Photometric and spectroscopic evolution of supernova SN ......Type Ia supernova, SN 2009an, over the...